Análisis situacional del sujeto de estudio

This section is about the several options we have for lead time reduction. From our analysis in chapter 2 and literature in chapter 3 we know the following ways for lead time reduction. We address the pros and cons of each alternative.

5.1.1

FORECAST

At first we have seen rolling forecast is in the starting phase for certain radar systems. This option can reduce lead times but will not necessarily do this. Depending on the timing of an order, and the number of previous orders, the lead time for the customer can be a few to X months. Currently at Thales the chance of getting an order should be very high before already start producing on this forecast.

Commitments with suppliers should be made about compliance of the forecasts, e.g. what if the forecast was wrong, if Thales does not sell the system, for what part of what already is produced is Thales accountable? If there is agreement for these kind of issues, this option can be promising. The commitment model proposed by (Fontijn, 2014) and (Thales, 2010), and this should be used for this option. We explained this model in section 3.6.

Forecasts in general are hard because radars have irregular and low demand. Also they are very expensive.

Therefore Thales only wants to start producing a radar when a customer has signed to purchase the radar, build-to- order. The current start-up at Thales for forecast on several systems which have items in common is an interesting direction, because this will reduce the chance that one of those items will not be sold.

After all forecast is not really an option for radar systems, although it would directly reduce the lead time because production can already start when the contract is not signed yet. The cost of eventually not selling the system are too high. For radars with a certain stable demand of more than two a year, the risks of eventually not selling the system are much lower. For this series production, forecast is definitely a way to reduce the lead time of these radars. This idea will develop at Thales the next years, but forecast will not be taken into account in this research. We assume the build-to-order policy, so the production starts when the customer signs the contract.

5.1.2

PRE-RELEASE OF LLI

The order of a system is officially signed by the customer, that moment is called the effective date of contract (EDC). Before this moment, some Long Lead Items that directly influence the total lead time can already start production. This is called a pre-release.

If the chance of getting an order is above a certain level, estimated by Thales, this option should be considered. This option can and is already used in the design phase of a system, for the first product, the first of class (FOC). But

36

it can also be used for products that already have been produced before. For this the critical paths should be determined and the paths that are longer than the lead time given to the customer, should be started before EDC. It should be started the number of days before EDC, such that the system is fully produced as contracted with the customer. However, the time of EDC is in this point in time just an estimation. The sales people should, next to the chance of getting the order, also give the expected time before EDC.

This possibility of course has risks, because the customer has not signed yet. Therefore the cost of eventually not using the Long Lead Item for this radar should be used in the trade-off. Long Lead Items used in multiple systems will have less risk, because perhaps they can be used later on for another radar.

5.1.3

TOWARDS STANDARD DESIGN

Uncertainty in lead time for items is higher for non-standard items. Relative long lead times are also for a huge part the result of being out of the standard. Therefore items with more uncertain lead times should have more

attention like more often updating lead times, and monitoring the progress of the item when ordered.

Manufacturers like to produce massive standard products, because that is where they can earn money the most. Thales orders are in a very low quantity and most of the time way more complex than the standard. Is Thales interesting for suppliers? Given the fact that suppliers cannot earn a lot of money from them, Thales is not interesting for suppliers. The economical situation is important in here. If we live in a recession, suppliers do not work with full production capacity, then Thales is more interesting for them. If the economy is booming, suppliers in general have enough work, then Thales with its relative complex and low quantity orders, is not interesting. If Thales could go with the flow of standard products, these above issues will not be the case. In the design phase of new systems, the standard of each critical item should be considered against the not standard option. An example is the use of 3mm aluminium, not standard. 1mm and 2mm are standard, gluing them together seemed no problem, dramatically reducing the lead time and cost. Therefore now the standard option is used, deleting all the uncertainty for this part.

Of course not every critical item can be standardized. Thales has to stay innovative and highly advanced. But for the parts where standardization is possible and is not making the system less valued, this option should be considered. Both for lead time and cost this can make a difference. Measures that have to be taken otherwise for lead time reduction, can be saved. The estimation is that in general for 5-10% of the items of a radar no standard option will be possible (Designer Thales, private communication).

GOOD ENOUGH VERSUS TOP LEVEL PERFORMANCE

Thales design engineers try to make the most technological sophisticated products, they want the best

performance for their customers, most of the time better than the customer requested. This fact gives Thales a very good impression to their customers, on performance of their products at least, but there are other

consequences. It results in a more complex design, which is often followed by a longer lead time and higher cost. The level of sophistication should be a trade-off with other important issues, like lead time for example. Cost is taken into account, but lead time is not. Because lead time is an important issue for management, we think it should have a larger weight in the trade-off with sophistication. This can result in less advanced systems, but if it is good enough for the customer, this is an arguable option. The customer also values the lead time criterion. Lead time should at least be considered in this trade-off, and the easiest improvement in lead time can be done in the design phase, otherwise we have to do symptoms treatment.

37

Applying this option on the whole radar will take a lot effort. However this option, like every other option, should be applied on the critical items of a radar. We will see in section 5.3 these are 39 items.

EARLY SUPPLIER INVOLVEMENT

In the design phase there is already cooperation between design engineers and suppliers, but this is more about manufacturability than the associated lead times. These conversations would be the right place for taking into account lead time issues. Lead times are not the only point these conversations should be about, but they would be a good addition to these conversations in the design phase, they can reduce critical long paths from the system. We already have seen the largest improvements can be made with purchased parts. Cooperation with the suppliers of those items in the design phase about lead times can prevent lead time issues to be solved later.

5.1.4

CUSTOMER ORDER DECOUPLING POINT

Products for which now a rolling forecast is started, all have in common a relative late CODP. Products fall into this category if they are fully developed, if they have a certain stable demand and the sales are at least two to four a year. The characteristic of the CODP is that before this point, there can be produced on forecast. After this point, the product is customer specific and only if the customer has given his specifications, the production can continue. These products have a make-to-order policy in the sense of rolling forecast, but due to the fact the order is not signed yet, it can be seen as make-to-stock. A late CODP seems to be a necessity for a shorter lead time by producing on forecast.

When designing a new systems however, having a late CODP should be taken into account as a goal. For lead time issues, this is recommendable. This can be achieved by designing the customisation of the system as late as possible in production. However, radars already have in general a CODP only a few months before finishing the system. And because radars are almost only build-to-order, the customer order decoupling point does not deserve that much attention.

5.1.5

INVENTORY

One way to reduce the lead time would be having every item, or a lot, in inventory. This would dramatically reduce the lead time, but is not very cost effective. The inventory cost should be used in the trade-off between lead time reduction and cost. This can be done with our model, like we explained in section 4.1. Earlier research at Thales (Thales, 2010) has been done on calculating the necessary inventory and accessory cost for certain lead time reduction of two systems. Eventually, nothing is done with this research, but the researchers saw for sure potential in inventory holding for improving lead times.

The anonymous stock which is kept standard in inventory is not relevant to this research, we assume these items are always available. The customer specific items are what we focus on in this research.

Customer specific items are items belonging to a certain system, or expensive items which are not standard in inventory. In the model we made, the necessary reduction of each item is determined, in order to achieve the goal set for the total lead time for the radar. With this information we can calculate the inventory cost per year for each item, resulting in total in the inventory cost for this lead time goal. We use 15% inventory cost per year.

It can be the case that just a small or cheap sub item of the item under investigation, is the cause for most of the lead time for that item. With this information that should be gathered about the item at suppliers, such affordable

38

and lead time effective measures can be taken. Inventory for sub-items reducing the lead time from a Long Lead Item is what we should look for.